| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Department of Internal Medicine, Yale University School of Medicine, New Haven, Connecticut;
Rega Institute for Medical Research, University of Leuven, Belgium; and
Centre d’Ecologie Cellulaire, Institut National de la Santé et de la Recherche Médicale, Hôpital de la Salpétrière, Paris, France
Correspondence: Dr. Stephen E. Malawista, Department of Internal Medicine, Yale University School of Medicine, New Haven, CT 06520-8031. E-mail: stephen.malawista{at}yale.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Key Words: locomotion • chemotaxis • chemokines
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Two of their observations were unheard of at the time: first, that leukocytes were able to locomote at all in the presence of EDTA, when the divalent cations necessary for the function of molecules of adhesion (integrins) would be largely sequestered [2 ], and second, that this unusual clustering about monocytes could occur. In subsequent work, it has become clear how the PMN were able to exhibit random locomotion and chemotaxis in EDTA. For maximal optical effect, the cells were somewhat compressed between slide and coverslip, and under these conditions, they could generate the force for locomotion by "chimneying"—i.e., by using the two surfaces as does a rock climber in a narrow crevasse or chimney—thus minimizing the need for molecules of adhesion [3 , 4 ]. The second observation, the clustering of PMN about monocytes and its nature, is the subject of this study.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Chemokines
Interleukin-8 (IL-8; #75500; homogeneous natural IL-8; 2000 ng/ml) was a mixture of the 77 amino acid (AVLPR) form (80%) and the 72 amino acid (SAKEL) form (20%) [5
], purified by controlled pore glass, heparin-Sepharose, fast protein liquid chromatography, and high-pressure liquid chromatography [6
, 7
]. Its cellular source was cytokine (IL-1)-stimulated osteosarcoma cells (MG-63). In a Boyden chamber assay, the chemotactic activity of this preparation was still significant at 10 ng/ml.
Human neutrophil-activating peptide-2 (NAP-2; #70900; homogeneous natural NAP-2; AELR; 2500 ng/ml) was purified from leukocyte-conditioned medium as for IL-8 [7 ]. Pure NAP-2 is at least 10 times less active chemotactically than the most active (SAKEL) form of IL-8 [8 ]. IL-8 and NAP-2 were frozen in aliquots until use.
Chemokine receptor blockers
These included the leukotriene B4 (LTB4) receptor antagonist CP105696 (Pfizer, Groton, CT), platelet-activating factor (PAF) receptor antagonists WEB2170 and 2086 (Boehringer-Ingelheim, Norwalk, CT), and CXCR2 [IL-8 receptor B (IL-8RB)] antagonist SKF83589 (SmithKline Beecham, King of Prussia, PA) in stock solutions of dimethyl sulfoxide (DMSO), usually at 10 or 20 mM, and blocking antibodies directed against CXCR1 (anti-IL-8RA; 9H1; J. Kim, Genentech, South San Francisco, CA) and CXCR2 (anti-IL-8RB; 10H2; Genentech) and against the C5a receptor (anti-C5ar; T. Hugli/J.Ember, Scripps, La Jolla, CA).
Enzyme inhibitors
We used the serine protease inhibitors benzamidine (Sigma Chemical Co.) and AEBSF (aminoethylbenzenesulfonyl fluoride, HCl; Calbiochem, San Diego, CA) and the metalloproteinase inhibitor phosphoramidon (Calbiochem).
Leukocytes
Fresh heparinized blood from human donors was allowed to sediment in tubes at an angle of 
60°C at room temperature for 1 h. Leukocytes from the buffy coat were concentrated in a microcentrifuge (Costar Corp., Cambridge, MA; Model #8455) for 30 s at 10,000 rpm (5585 g) and were resuspended in 1 ml autologous heparinized plasma (except as noted), along with a small number of added erythrocytes. Reagents were diluted in PBS before their final addition to the leukocyte suspensions, resulting in final plasma concentrations of between 50% (distraction experiments with NAP-2 and IL-8) and 85% (experiments with multiple cytokine receptor antagonists), with corresponding control preparations containing the same dilutions of PBS, DMSO, or nonimmune serum.
A drop of this suspension sufficient only to wet an entire overlying 22 mm x 32 mm coverslip (4 µl) was deposited on a clean glass slide, and the preparation was sealed with paraffin and removed to the warmed (37°C) stage of a Zeiss phase-contrast photomicroscope (objective, 40x or 25x, except as noted), connected via a Hamamatsu microscope video camera C2400 (Hamamatsu Photonics K. K., Hamamatsu City, Japan) to a Panasonic time-lapse video recorder AG6720 (Matsushita Electric Industrial Co., Osaka, Japan).
Chemotaxis
The orientation and trajectory of leukocytes were observed and recorded in time-lapse video microscopy (16x real time). In addition, chemotactic gradients lasting for many minutes were created by the destruction of an erythrocyte or adjacent ones, by a ruby laser microbeam (wavelength, 694.3 min; duration of flash, 0.5 ms) focused backward through the optics of the microscope to a diameter of 5 µm [9
]. The nature of this chemoattractant is unknown. Again, the orientation and trajectory of every cell in the field were recorded [10
].
Handling qualitative data
Before the start of videomicroscopy of a given preparation, the slides were examined at low power for aggregations of PMN about monocytes. The controls used in EDTA/plasma typically develop hundreds of these clusters over many minutes. The number of cells per triggering monocyte is variable within a given preparation.
When inhibitors were used, they were pushed to concentrations that essentially abrogated the development of such aggregates—meaning none or an occasional, somewhat chemoattractive monocyte. To show that the PMN in this preparation were still capable of chemotaxis and not simply injured, the alternative chemotattractant, the irradiated erythrocyte, was then used in the same sealed preparation. Again, these were not ID50s, but qualitative differences seen in multiple preparations; the effect is there or not. Fisher’s Exact Test was used (see Table 1 ) when there was biologic variation among donors in these essentially all-or-none phenomena.
|
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
Concentrations of EDTA
The effects of EDTA/plasma were first noted by one of us (A. d. B. C.) using a standard anticoagulant preparation of disodium EDTA (Sigma Chemical Co.), 1 drop/ml plasma (1.8 mM). Because in the blood of some donors the effect was often not seen until that concentration was increased, we gradually used higher concentrations and for purity and precision, moved to highest-grade, buffered EDTA (and later, EGTA; see Materials and Methods). Eventually, we chose a standard concentration of EDTA, 10 mM, in which the cells appeared healthy, in which cells from the large majority of donors responded, and which would have decreased the concentrations of free Ca++ and Mg++ to 
1 nM.
Chemoattraction of PMN by monocytes and the licking response could also be seen in CDTA/plasma (see below), as well as in citrated plasma, but in the latter only at very high concentrations (50–100 mM). Heparin sodium alone had no effect on the chemotactic response to monocytes, but we eventually traced a lack of effect of added EDTA in some bloods to their initial anticoagulation with heparin lithium, which we stopped using.
Chelation of Ca++ alone was not sufficient for these effects to be seen. EGTA in plasma at concentrations as high as 10 mM did not trigger chemoattraction of PMN by monocytes. It did induce the segregation of PMN into groups, but not about monocytes and without the licking response. Early in this study, when we were gradually increasing the concentration of EDTA to see where its effect became apparent for a given donor, we sometimes saw similar aggregation of PMN at suboptimal concentrations of EDTA. This picture could also be seen when MgSO4 was added to EDTA/plasma. Thus, chelation of Ca++ and Mg++ was necessary for the effect seen in Figure 1 .
Monocytes: the importance of adherence
Typically, a great many monocytes in a given EDTA/plasma preparation recruit PMN, but many others do not. A priori, the former cannot be distinguished microscopically from the latter, but adherence of an attracting monocyte, however tenuous in EDTA, seemed important. This was indicated when a monocyte gave up its central position and began locomoting (Fig. 2
) or sometimes floating, dislodged by the vigorous licking or for some reason unperceived. PMN generally ignored it as it passed through their ranks; their attention remained focused on the place where it used to be.
|
|
|
When only Ca++ was chelated—i.e., in EGTA/plasma—PMN responded normally to a necrotactic stimulus (not shown), indicating their general health. Although they sometimes took abnormally long to disperse, they did not exhibit a second wave of chemoattraction. The lack of a second wave suggests that their chemoattractant range was not great or was not prolonged to the extent that it was in EDTA/plasma.
Identification of the monocyte chemoattractant: attempts to "distract" PMN from chemoattractant gradients in EDTA/plasma by addition of putative key chemoattractants
Our hypothesis was that the monocyte chemoattractant was likely to be IL-8 or NAP-2. To pursue this possibility, we added these cytokines, alone or in combination, to slide preparations to determine whether we could compete with chemoattractant gradients in which they might play a role. That is, by adding the same factor that was present in the gradient, we sought to decrease the steepness of the gradient to such an extent that the recruitment of PMN was prevented.
For each experiment described, there was good clustering of PMN about monocytes in controls containing EDTA, 10 mM/plasma; control clusters are easily seen (see, e.g., Fig. 5D ). In contrast, as shown in Table 1 , column 3, in the presence of added NAP-2, 500 ng/ml, chemoattracting monocytes were seen in only one of six separate experiments (P=0.0076). If one includes multiple, separate preparations from the same donor, the figure becomes 1 of 11 preparations (P=0.00025). [In all preparations that received NAP-2 (columns 3+5), the figure is 4 of 19 (P=0.0001).] In contrast, clusters of PMN about monocytes were found in four of the first five preparations to which IL-8, 400 ng/ml, was added alone (experiments 1, 2, and 5) or with NAP-2 (experiment 3).
|
Clustering about monocytes may become less when EDTA is added after the cells have been resting for several hours at room temperature. Therefore, we repeated EDTA/plasma controls in all preparations in which the last experimental group had no clusters about monocytes. Late clusters were present in all such experiments.
Thus, in these experiments aimed at distracting PMN from chemotactic gradients, we take the results with exogenous NAP-2 and with exogenous NAP-2 plus IL-8 as evidence that NAP-2 may be the chemokine generated by monocytes in EDTA/plasma.
The importance of platelets, which provide the precursors of NAP-2, for the initial clustering of PMN about monocytes in EDTA/plasma
To pursue the possibility that NAP-2 was the monocyte chemoattractant, we next examined whether platelets appeared to be involved. NAP-2 is generated by monocytes but not synthesized by them; they make NAP-2 by the final processing of thromboglobulin, a peptide precursor elaborated by platelets [13
, 14
].
We therefore washed leukocytes from heparinized blood five times in PBS to remove as many adherent platelets as possible, and removed platelets from half the autologous plasma by rapid microcentrifugation. We then added plasma ± platelets and EDTA, 10 mM, to the washed cells and counted the number of aggregates that formed in the paired preparations. Here, we were looking for quantitative differences, because some platelets continue to sediment with washed leukocytes, and precursors of NAP-2 already released in plasma would not sediment with the platelets.
The results are seen in Table 2 . In five experiments, there was a mean of five times as many aggregates when plasma not depleted of platelets was used, and the paired differences were highly significant (P<.01).
We take this as further evidence for a role of NAP-2 in the initial response of PMN to monocytes in EDTA/plasma.
Effect of blocking the specific receptor on PMN that NAP-2 uses
If NAP-2 were the chemoattractant for PMN that was being generated by monocytes in EDTA/plasma, then we reasoned that blocking its receptor, CXCR2 (previously referred to as IL-8RB), would inhibit the effect. It is interesting that our antibody to CXCR2 and the CXCR2 receptor blocker SKF83589 acted in a complementary manner: The former (10 µg/ml) combined with 10 µM of the latter resulted in little or no licking of monocytes by PMN. The antibody by itself did not suppress this reaction in concentrations between 10 and 400 µg/ml (the highest concentration used), but the SKF compound alone did abrogate it at 100 µM (Fig. 5A
). Thus, the specificity of the two blockers appears to be overlapping but not identical. Neither of these materials affected the general chemotactic function of the PMN, as evidenced by the persistence of necrotaxis and the development of a second wave in EDTA/plasma (Fig. 5A
, inset). In contrast to the effect of agents that block CXCR2, chemoattraction by monocytes in EDTA/plasma was not affected by the antagonists of LTB4 or PAF receptors, 10–50 µM, or by blocking antibodies of CXCR1 (IL-8RA) or of the C5ar, 10–50 µg/ml; preparations with none of these other reagents could be distinguished from EDTA/plasma controls (Fig. 5B)
. Nor was the attraction of PMN to monocytes in EDTA impeded by the use of autologous serum heated to 56°C for 30 min (to inactivate complement) instead of plasma.
Effect of inhibiting serine-proteases necessary for the generation of NAP-2
Monocytes are thought to generate NAP-2 from its platelet-derived peptide precursor by elaborating cathepsin B or other serine proteases [14
]. Initially, we compared the effect of the serine-protease inhibitor, benzamidine, to that of the metalloprotease inhibitor, phosphoramidon, or to that of EDTA alone in this system. In several experiments, benzamidine, usually 4 mM, added to plasma and cells at the same time as EDTA, usually 10 mM, resulted in much smaller numbers of PMN about monocytes, or none, and very few such groupings compared with simultaneous or subsequent EDTA controls. However, necrotaxis was normal, and in one experiment, PMN grouped about detritus on the slide, showing that this was not a nonspecific toxic effect; the PMN were capable of responding normally to other chemotactic stimuli. Phosphoramidon, 40 µM, had no effect.
We confirmed this finding with another serine protease inhibitor, the sulfonylfluoride, AEBSF. In several experiments, AEBSF, 0.2–0.25 mM, plus EDTA, 10 mM, prevented or dramatically reduced the chemoattraction of PMN for monocytes in plasma (Fig. 5C) compared with simultaneous EDTA/plasma controls (Fig. 5D) . AEBSF did not affect the general chemotactic function of the PMN, as evidenced by necrotaxis and the development of a second wave in EDTA/plasma (Fig. 5C , inset). We take these findings as further evidence that NAP-2 is the critical early chemoattractant toward monocytes.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The site of generation of NAP-2
NAP-2 was discovered in cocultures of monocytes and platelets [13
, 15
]. In the current study, the geography of its generation is particularly well demonstrated. That PMN directly address the monocyte (Fig. 1)
indicates that NAP-2 is generated very near the monocyte cell surface; i.e., the platelet-derived material is presented to the monocyte and not the other way around. If processing took place as the precursor peptides left the platelet, PMN would cleave to platelets; if it took place throughout the preparation, there would be no gradient. Hence, the behavior of the PMN marks the site of generation of the chemoattractant.
Analysis of apparent site specificity of the initial chemotactic response of PMN to monocytes
In EDTA/plasma, the monocyte is not squeezed out [1
] by the PMN that it attracts. Rather, it leaves the field in good morphologic health after becoming less adhesive and losing its attractive power (Fig. 2)
. The PMN do not follow the migrating monocyte; they ignore it, remaining focused on the place where it used to be (Figs. 2
and 3)
. That place might be occupied, transiently, by a PMN (Fig. 2)
or by successive ones, but eventually there was often no designatable central cell at all (Fig. 3B)
.
How does one interpret these findings? For example, could the targets be acting as "salt licks" in an environment containing insufficient divalent cations? Assuredly not. Cells are not functionally depleted by their tenure as targets—PMN and often monocytes remain actively motile—and sometimes, there were no cellular targets. Moreover, we saw the same effects in 10 mM EDTA, where concentrations of free calcium and magnesium are less than nanomolar, so that there is no possibility of a divalent-cation gradient. Alternatively, could some hypothetical material deposited on the glass trigger the secretion of chemokines by successive cells that encounter it? Possibly, but it could hardly maintain a chemotactic gradient when there was no central cell at all. Taken together with the eventual large size of many of these aggregates (Fig. 3B) —i.e., the distance of peripheral PMN from central ones—we suspected that the recruited PMN in EDTA/plasma had themselves taken over the generation of a chemotactic gradient, replacing the one initially provided by the adherent monocyte.
Necrotaxis: a "second wave" of chemoattraction in EDTA/plasma
Supporting this hypothesis were the findings in another chemotactic gradient that did not involve monocytes: the gradient created by an erythrocyte destroyed by laser microirradiation (necrotaxis; Fig. 4
) [9
]. Instead of dissipating after several minutes, the PMN in EDTA/plasma kept coming, forming very large aggregates about the target area. We refer to this phenomenon as a "second wave" of PMN, although it is properly a continuation of the initial response. Again, it appears that the PMN in EDTA/plasma have taken over the generation of a chemotactic gradient, replacing in this case the one initially provided by the irradiated erythrocyte.
Prolonged chemoattraction in EDTA/plasma by PMN activated in various ways
Thus, the early monocyte response seems to be simply one way of stimulating the PMN, which once activated, fail in EDTA/plasma to shut off their own chemoattraction for other PMN. Other examples of this prolonged chemoattraction by activated PMN in EDTA/plasma would appear to include the large aggregates that accrue about a PMN containing ingested material, about an occasional "sick" or dead (motionless, mummified-appearing) PMN, or about detritus in the slide preparation. The only other cell type that seemed to initiate chemoattraction of PMN in EDTA was the occasional metamyelocyte seen in these preparations; their appearance was too rare for us to pursue the source of their chemoattraction.
Hypothesis: tonic inhibitor(s) of chemotaxis (TIC) in plasma
We appear to have uncovered a new EDTA-inhibitable, regulatory mechanism for chemotaxis in plasma and perhaps elsewhere as well. Specifically, we suggest that these exaggerated chemotactic effects are due to the loss of normal modulation by a regulatory factor(s) designed to keep the chemotactic response from getting out of hand—i.e., that this hypothetical modulator(s) is a tonic inhibitor of chemotaxis (TIC) in plasma. To operate efficiently, such a hypothetical modulator would likely come from the same cells that are producing chemokines; modulation at a distance is not likely to work. When the initial inflammatory stimuli have been overcome, modulatory elements would be necessary to dampen cell-cell recruitment. An early example of same-cell regulation, perhaps to keep responses within limits, is the generation of C5a by a specific (secondary) granule product (the granules more readily secreted) and inactivation of C5a by azurophil primary granule products [16
].
Again, although the dramatic aggregation of PMN about monocytes in EDTA plasma (Fig. 1) was what piqued our interest initially, and its analysis is the primary subject of this paper, monocytes per se are not central to this hypothesis. Rather, anything that will activate PMN is a candidate for breaking through the tonic inhibition.
Teleologically, this formulation makes a good deal of sense. Recruitment of PMN cannot be allowed to continue once a bacterial invader, for example, has been disposed of; modulatory elements have to get the upper hand. It is likely that vascular mini-traumas are frequent, activating cells to produce chemokines, and it is essential that the response to "false alarms" not get out of control. Reassertion of some baseline level of tonic inhibition would seem more efficient than simply waiting for such stimuli to dissipate. In this paradigm, chelation of divalent cations by EDTA may reduce the ambient levels of the tonic inhibitor(s), by inhibiting its production or secretion or by accelerating its rate of breakdown. The result is a prolonged ingress of PMN—the second wave. If a TIC indeed exists and is obtainable (a subject of current interest), it would clearly have potential therapeutic, anti-inflammatory usefulness.
Licking as retrograde flow
In EDTA/plasma, PMN locomoting randomly or responding chemotactically in thin preparations are morphologically similar to untreated PMN. Only on their arrival at a target do they exhibit the exaggerated size and activity of protopods that in time-lapse videomicroscopy resembles vigorous licking. Thus, something changes on arrival. What follows is largely speculative, but it perhaps puts in perspective what we are observing.
We were able to create targets in EDTA/plasma in which the protopods of PMN were even longer than those seen here (unpublished results). In time lapse (and with a speed that can even be appreciated in real time), one sees in phase-contrast microscopy a billowing backward of cytoplasm from the site of EDTA-frustrated phagocytosis toward the cell body, which can go on for hours. This process resembles a more rapid version of the cytoskeletal reorganizations that occur, for example, in neuronal growth cones in Aplysia (sea slug) bag cell neurons. Like protopods of PMN, growth cones may be viewed as signal transduction devices, which interpret extracellular signals and then physically direct neurite outgrowth (e.g., advance, withdrawal, turning, branching) through the regulation of intracellular, cytoskeletal dynamics and the molecular motors that interact with cytoskeletal proteins [17 ]. In both cell systems, F-actin networks are preferentially assembled at the leading margins, then flow centripetally, and are disassembled and recycled. In Aplysia, the advance of the growth cone is inversely proportional to retrograde F-actin flow within them [18 ], and it seems to result from two independent processes: actin assembly and myosin-based filament retraction [19 ]. This may be a general strategy for locomotion in motile cells [20 ].
Our hypothesis is that we are seeing similar events in PMN in EDTA/plasma: PMN approach their targets polymerizing actin at their advancing fronts, as do control cells, but having arrived and finding themselves unable to engulf the target efficiently in EDTA, they cycle the F-actin that they are throwing at this problem into retrograde flow and proximal disassembly. We see this as billowing backward from the site of interaction between protopod and target or in the prominent but shorter protopods in the current study, as vigorous licking.
Additional questions for future work
None of the reagents we used—the antagonists of CXCR2 and of receptors for LTB4, and PAF; the antibodies to CXCR1, CXCR2, and to the C5ar; or the serine protease inhibitors benzamidine and AEBSF—suppressed necrotaxis or the chemoattractant properties of activated PMN for other PMN. This was in fact fortunate, because if everything had stopped, we could not have ruled out toxic effects on cells, and the data would not have been interpretable. However, the question remains as to what these other attractants are. The answers may become evident as additional chemotactic receptors are cloned and as specific blockers for them are developed.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
ACKNOWLEDGEMENTS |
|---|
Received July 8, 2001; revised February 7, 2002; accepted February 28, 2002.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. E. Malawista, A. de Boisfleury Chevance, J. van Damme, and C. N. Serhan Tonic inhibition of chemotaxis in human plasma PNAS, November 18, 2008; 105(46): 17949 - 17954. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
